The rationale for the selection of hormonal versus chemotherapy for breast cancer patients is based on the biochemical parameter for hormone dependency, namely the presence of cytoplasmic estrogen receptors (ER). Tumors lacking ER and classified as hormone-independent rarely regress in response to hormonal therapies (additive or ablative). The frequency of "false negatives" in this group is only 10%. On the other hand, the probability of misclassification of ER-positive tumors is hormone-dependent and, therefore, as responders to hormone therapy is 40% if the presence of ER alone is considered relevant and 30% if the presence of ER and another hormone receptor, progesterone receptor (PR), is taken into account. Since ER-positive tumors are the more predominant kind (75% of all breast cancers are ER+) there is a need for greater accuracy in the selection of truly hormone-dependent tumors for hormonal therapy. In addition, cytoplasmic ER is also a marker protein for better prognosis, late recurrence pattern, and longer disease-free survival. We propose to examine the utility of classification of breast tumors based on immunohistochemical parameters either applied individually or in conjunction with the biochemical data on ER for more precise selection of hormone-dependent from hormone-indifferent tumors. Two different approaches to immunohistochemical classification will be compared for their relative merits: (a)\Classification based on the proportion of ER-positive and ER-negative tumor cell population. This will be achieved by analyzing frozen sections from multiple areas of the tumor for the density of ER-positive and ER-negative cells. ER-positive cells will be identified by immunofluorescent (IF) procedure using polyclonal antibodies specific for Type I ER; (b)\Transportation of E2via transformed ER to specific nuclear acceptor sites is a prerequisite for all the molecular events (including PR synthesis) that are dependent on estrogen. Using immunohistochemical procedures ER-positive tumors will be sub-classified on the basis of the density of ER-positive cells that can translocate their ER, in vitro, when exposed to estrogenic or antiestrogenic ligands. The efficacy of this mode of classification for selecting the hormone-dependent tumors will be compared with that of the biochemical mode of classification based on the presence/absence of PR for selecting the hormone responders in the ER-positive group. (2)